DM Chapter7

7/31/2019 DM Chapter7

1/14

Chapter7Diabetes andStroke

7.151

Diabetes in Ontario Practice Atlas

Authors: Moira K. Kapral, Deanna M. Rothwell, Kinwah Fung,Mei Tang, Gillian L. Booth and Andreas Laupacis


2/14

BackgroundStroke is a leading cause of death and disability in Canada.1 Diabetes

mellitus (DM) increases the risk of stroke, and is a particularly

potent stroke risk factor in younger individuals, with previous

studies suggesting an increase in stroke risk of as much as 10-fold

in some younger subgroups.2,3,4 DM is associated with a higher

prevalence of other stroke risk factors, including high bloodpressure and high cholesterol,5 and may increase the risk of stroke

recurrence and mortality.6,7 Despite the association between DM

and stroke, the available data from clinical trials do not support the

hypothesis that better blood sugar control decreases stroke risk.6,8

Carotid endarterectomy is a surgical procedure to remove athero-

sclerotic plaque from the carotid artery. Clinical trials have found

that in appropriately selected individuals with previous stroke or

transient ischemic attack (TIA), carotid endarterectomy substantially

lowers the risk of future stroke or death compared to medical

therapy.9-11 It is not known whether carotid endarterectomy rates

are different in those with and without DM. Perioperativecomplications may be more frequent in those with DM.12

This chapter will present analyses of stroke-related hospitalizations,

outcomes (death, length of stay and discharge to complex continuing

care institutions) and procedures (carotid endarterectomy) in Ontario,

in people with and without DM, with stratification by age, sex,

socioeconomic status and geographic region.

Data SourcesThe Registered Persons Database (RPDB) was used to identify all

individuals between the ages of 20 and 105 who were eligible for

coverage under the Ontario Health Insurance Plan (OHIP) during the

fiscal years 1995 to 1999. Persons with DM were identified using the

Ontario Diabetes Database (ODD), which is described in detail in the

Chapter 1 Technical Appendix TA1.A. Individuals in the RPDB who were

not present in the ODD served as a non-diabetic comparison group.

Creation of this cohort is described in Chapter 5 Technical Appendix

TA5.A. Records of hospitalizations for stroke and carotid endart-

erectomy procedures were obtained from the Canadian Institute for

Health Information (CIHI) discharge abstract database. Census data from

Statistics Canada were used to obtain information on the socio-

economic status of residential neighbourhoods. These data were linkedto other sources using postal code of residence as a common variable.

How the analysiswas doneAnnual stroke hospitalization rates were calculated from fiscal 1995

(April 1, 1994 to March 31, 1995) through fiscal 1999. The total

number of persons with DM who were admitted with a stroke in

a given year defined the numerator, while the denominator was the

total number of persons with DM during the same time period.

7.152

Diabetes in Ontario

Key Messages

Stroke risk is markedly increased in thepresence of diabetes mellitus (DM), even in

younger individuals. Health care professionals

and patients should be aware of these risks,

and should be attentive to the appropriate

management of associated stroke risk factors.

Stroke hospitalization rates in those with DMare declining over time. Further research is

needed to determine whether this is due to

changes in stroke incidence or stroke admission

thresholds.

Diabetes and Stroke


3/14

Records of admissions to acute care hospitals with a diagnosis of

stroke were obtained from the CIHI discharge abstract database

using ICD-9 codes 431, 434 and 436. Previous studies using these

codes have established an accuracy rate of over 90% for the

diagnosis of stroke.13 Persons transferred from other facilities and

those with stroke as an in-hospital complication

were excluded from the analyses. For those

persons with more than one stroke admission

during the study time frame, only the first

stroke admission was included in the analyses.

Annual stroke hospitalization rates were

calculated for persons with and without DM, and

were categorized by age, sex, socioeconomic

status, and geographic region (county). In

Ontario, personal income is not available

in administrative data sources. Therefore,

neighbourhood median household income

was attributed to the individuals studied.

Neighbourhoood level income quintiles were

obtained from 1996 census data at the level of

the enumeration area.14 This method defines

quintiles separately for census metropolitan

areas (CMA) or census agglomerations (CA)

and areas not in any CMA or CA, so that the

measure is relative to the larger area in which

a person resides.

Among stroke patients, median length of stay

and rates of discharge to complex continuing care

institutions were compared in persons with and

Practice Atlas

7.153

Stroke hospitalization rates are almost three-fold higher in individuals with DM compared to those without DM, and therelative increase in stroke risk is particularly marked in the younger age groups.

1995 DM 147.8 575.8 1,427.5 2,846.8 177.3

1996

DiabeticStatus

Women by Age Group

DM 109.9

738.0 1,661.5

541.9 1,323.7

3,290.8

2,800.1

Men by Age Group

177.2 743.3 1,700.7 3,077.3

136.2 691.4 1,639.9 3,073.2

14.4 156.0 574.0 1,585.8

14.4 147.0 590.0 1,639.9

1997 DM 116.4 579.6 1,303.9 2,706.4

No DM 12.6 86.9 372.1 1,334.3

No DM 11.6 94.6 373.9 1,344.5

Source: Ontario Diabetes Database (ODD). *Odds Ratios (95% CI) are only for 1999. Note: Fiscal year 1995 = April 1, 1994 to March 31, 1995. Adjusted for age and

Exhibit 7.1 Overall and Age-/Sex-specific Stroke Hospitalization Rates per 100,000 Ontarians with/without DM19951999

4,526

4,792

5,071

12,078

11,972

No DM 11.1 88.8 395.6 1,374.712,131 14.1 154.5 639.1 1,691.7

1,214

Overall Men &Women

1,177

158

1,151

154

154

5,125

11,906

152.3 605.0 1,526.9 2,883.8

15.2 142.2 548.7 1,527.3

1998 DM 135.3 513.6 1,243.3 2,509.01,074

150No DM 11.9 80.6 347.2 1,334.5

5,222

11,356

164.4 594.8 1,405.3 2,688.7

14.0 132.2 514.4 1,435.6

1999 DM 82.5 467.4 1,130.9 2,487.51,015

141No DM 10.5 79.3 349.1 1,218.6

1,203

1,145

156

1,134

154

155

1,059

151

999

142

1,225

1,207

1,167

153

154

160

1,088

148

1,029

139

OverallnFiscal Year 2049 5064 6574 75+ Overa2049 5064 6574 75+Rate

7.

(7.167.

11.74

(9.4014.65)

4.52

(4.085.01)

2.76

(2.542.99)

1.90

(1.772.04)7.28 (7.057.53)

Odds RatioCrude*

7.08

(6.757.42)

7.87

(5.7810.72)

5.92

(5.196.75)

3.27

(2.983.5)

2.07

(1.942.20)

2.

(2.542.2.67 (2.582.76)

Odds RatioAdjusted*

2.67

(2.552.81)

Exhibit 7.2 Age-/Sex-specific Hospitalization Rates for Stroke per

100,000 Ontarians with/without DM Aged 20 Years and Over, 1995199

There was a decline in stroke hospitalization rates over the study periodin persons with and without DM.

Source: Ontario Diabetes Database (ODD). Note: Fiscal year 1995 = April 1, 1994 to March 31, 199


4/14

without DM, with stratification by stroke type (hemorrhagic and

ischemic). Mortality rates following hospitalization for stroke were

calculated at 30 days and one year after the index admission.

Deaths were ascertained from the Registered Persons Data Base

(RPDB) and CIHI discharge abstracts. Mortality rates after stroke

were adjusted for age, sex, stroke type and

comorbidity based on the Charlson-Deyo score,

a commonly used method that uses indicators of

major disease groups within hospital diagnostic

codes to assign a level of comorbidity.15

Admissions for carotid endarterectomy pro-

cedures within a year after the index admission

were identified from CIHI records in which the

Canadian Classification of Procedures (CCP) code

was 50.12. Carotid endarterectomy rates per

100 patients with stroke were calculated for

persons with and without DM, categorized by

age and sex. Waiting times for carotid surgery

were calculated using the time from stroke

admission to the date of surgery.

Multivariate techniques (Cox proportional

hazards models) were used to identify risk

factors for suffering a stroke during the five-

year observation period. Factors that were

tested included age, sex, socioeconomic status

(SES), presence of other medical conditions

(comorbidity), type of residential area (urban

versus rural), geographic region of the province,

and use of outpatient services. Individuals were

categorized as having a regular provider of care

Diabetes in Ontario

7.154

Source: Ontario Diabetes Database (ODD)

Diabetes and Stroke

Stroke hospitalization rates in both those with and without DM were inversely related to socioeconomic status, with modestlyincreased stroke hospitalization rates seen in the lowest income quintiles.

Exhibit 7.3 Overall and Five-year Average Age-/Sex-specific Stroke Hospitalization Rates per 100,000Ontarians with/without DM by DM Status and Residence Area Income Quintile, 19951999

1,141Q1 (low) DM 145.9 636.4 1,405.8 2,586.81,215

1,078

1,194164.9

1,111Q2

Diabetic

Status

Women by Age Group

DM 124.9

779.1 1,736.7

521.4

938

2,275

2,472

176

1,304.0

No DM 13.0

3,131.8

118.3

2,703.5 1,172

Men by Age Group

150.7 706.9

431.5 1,384.32,678

1,638.4

1,134

16.3

3,018.0

149

155.2 682.0 1,487.9 3,078.9

159

15.8 145.1 566.2 1,517.6

164

14.7 168.6 606.0 1,588.4

196.3 679.6 1,753.7

1,166

Overall Men &

Women

1,045Q3 DM 91.8 501.9 1,179.4 2,679.3

1,143

170

1,092

148

159

147No DM 11.7 92.9 360.9 1,304.9

804

1,967

1,085179.1 609.3

159No DM 12.5 85.7 380.2 1,355.0

1,572.2 2,965.0

13313.5 124.9 528.3 1,509.2

1,056Q4 DM 101.6 489.4 1,198.6 2,860.01,072

130 127No DM 11.1 77.1 343.0 1,309.7

707

2,029

1,053153.8 564.3 1,419.1 2,807.8

13811.4 110.6 478.5 1,503.4

1,035Q5 (high) DM 98.5 426.8 1,198.4 2,611.01,045

131 126No DM 9.2 62.1 303.5 1,241.1

OverallnIncomeQuintile

2049 5064 6574 75+ Overall2049 5064 6574 75+Rate

Exhibit 7.4 Five-year Average Stroke Hospitalization Ratesper 100,000 Ontarians with DM Aged 20 Years and Over by

Residence Area Income Quintile, 19951999Stroke hospitalization rates in both persons with/without DM wereinversely related to socioeconomic status.



5/14

if at least 50% of their primary care visits were to a single provider.

Adjustment for the presence of other medical conditions that

might affect outcomes was performed using the John Hopkins

Ambulatory Care Groups (ACG) assignment software.16.17 Region

of residence was based on the Ontario Ministry of Health and Long-

Term Care (MOHLTC) planning regions. There was no significant

colinearity between any of the variables included in the model.

Small area rate variation (SARV) analysis compared hospitalization

and procedure rates across regions of the province (a review of

SARV statistics appears in Chapter 2 Technical Appendix TA2.1).

Interpretive CautionsThe analyses rely on administrative data, which lack information

on important prognostic factors such as stroke subtype and severity.

In addition, comorbid conditions may be miscoded. The analyses

use neighbourhood income quintile as a marker of socioeconomic

status, rather than individual level data, which may lead to

misclassification of individual socioeconomic status. The CIHIdatabase only captures hospital admissions, which would lead to

under-reporting of stroke rates since it does not include patients

with stroke or transient ischemic attack who were never admitted

to hospital. It is not known whether any of these factors would

lead to systematic bias in comparisons between individuals with

and without DM. However, it is conceivable that a person with DM

and minor stroke or TIA is more likely to be admitted to hospital

than a person without DM with a similar stroke presentation. This

could lead to higher stroke admission rates (and decreased stroke

severity among admitted patients) in persons with DM relative to

those without.

Trends in the control of risk factors, especially hypertension, are likely

to be important determinants of stroke rates over time, and these

were not evaluated in the current analyses. For the analyses of

carotid endarterectomy rates, administrative data do not have

information on the prevalence or degree of carotid stenosis or the

indications for surgery, so one cannot comment on the

appropriateness of the observed rates of surgery in this study

population. In addition, waiting times for carotid endarterectomy

are estimated based on the time between the index stroke

admission and the date of surgery, and may not be an accurate

reflection of the time interval from diagnosis or referral to surgery.

Findings and DiscussionEven after adjustment for age and sex, stroke risk was greatly

increased in those with DM, with stroke hospitalization rates

almost three-fold higher in individuals with DM than in those

without (Exhibit 7.1). The diabetes-related stroke risk was particularly

marked in the younger age groups, such that their stroke risk was

similar to what would be expected in an older non-diabetic

population. For example, the risk of stroke in a 20 to 49-year-old

person with DM was greater than that of a 50 to 64-year-old

Practice Atlas

7.155

Key Research Findings

Stroke hospitalization rates are approximatelythree-fold higher in those with diabetes

mellitus (DM) compared to those without,and are even more markedly increased in

younger age groups.

Stroke hospitalization rates are decreasingover time.

There are only minor regional variations instroke hospitalization rates across Ontario.

Older age, male sex, lower neighbourhoodincome quintile, previous myocardial

infarction and comorbid illness are allassociated with increased stroke admission

rates; conversely, the presence of a regular

source of care and the number of ambulatory

care visits do not appear to affect stroke

admission rates.

After stroke admission, those with DM areat increased risk of death within 30 days or

discharge to chronic care compared to

those without DM.

Men are more likely than women to undergocarotid endarterectomy after stroke.


6/14

Diabetes in Ontario Diabetes and Stroke

7.156

Exhibit 7.5a Five-year Averaged Crude and Age-/Sex-adjusted Stroke Hospitalization Rates per 100,000Ontarians with DM Aged 20 Years and Over by County, 19951999

Regional differences in hospitalization rates for stroke were not statistically significant in individuals with DM.



7/14


8/14


7.158

Sources:Can

adianInstituteforHealthInformation(CIHI),

OntarioDiabetesDatabase(ODD)

Exhibit7.6a

Age-/Sex-adjustedStrokeHospitalizationRatespe

r100,000OntarianswithDM,Age

d20YearsandOver,byCounty,N

orthernOntario,

19951999

Note:

See

Exhibit7.6

b

forLegend.


9/14

Practice Atlas

7.159

Exhibit7.6b

Age-/Sex-adjust

edStrokeHospitalizationRatespe

r100,000OntarianswithDM,Age

d20YearsandOver,byCounty,S

outhernOntario,

19951999

Peterborough


10/14


Exhibit 7.7 Characteristics of Stroke Patients by DM Status in Ontario, 19951999

There were significant differences in

stroke type in persons with/without DM,

with a higher proportion of ischemic

stroke among those with diabetes.


7.160

Exhibit 7.8 Age-adjusted Rates of Discharge to Complex Continuing Care Institution or Death within

30 Days after Stroke per 100 Ontarians with Stroke by DM Status, Stroke Type, and Gender, 19951999

After admission to hospital for either ischemic or hemorrhagic stroke, mortality within 30 days or discharge to chronic carewas higher in those with DM.

Mean ageyears

Q1 (low) 4,952 (24%)

72.92

Malen (%) 11,450 (53%)

12,243 (23%)

73.81

27,335 (48%)

N 21,774 56,759

11,308 (21%)

Income quintile n (%)

Q2

9,337 (18%)

4,583 (23%)

9,517 (18%)

10,571 (20%)


11/147.161

Practice Atlas

Exhibit 7.9 Thirty-day and One-year Mortality Rates per 100 Persons with Stroke by DM Status andGender in Ontario, 19951999

After adjusting for age, sex, comorbid conditions and stroke type, there were no large differences in all-cause mortality at30 days or one year in those with or without DM.

Source: Ontario Diabetes Database (ODD). Note: Fiscal year 1995 = April 1, 1994 to March 31, 1995.

Exhibit 7.10 Sex-specific Post-stroke Carotid Endarterectomy Rates per 100 Ontarians with Stroke andWaiting Times by DM Status, 19951999

Overall rates of carotid endarterectomy after stroke were similar in those with and without DM.

1-year Mortality

Number of Cases Risk-adjusted Rate*

30-day Mortality

Gender/DM Status Number of Cases Risk-adjusted Rate*

13,131Overall

4,024 34.40 (95%CI; 33.59-35.21)

9,107 35.37 (95%CI; 34.82-35.93)

7,728

DM 2,295 20.63 (95%CI;19.90-21.36)

5,433 20.82 (95%CI;20.35-21.30)

Men

No DM

14,411Overall

3,725 33.58 (95%CI; 32.76-34.41)

10,686 35.62 (95%CI; 35.11-36.13)

8,580

DM 2,170 20.36 (95%CI;19.61-21.10)

6,410 21.91 (95%CI;20.47-21.34)

Women

No DM

Source: Ontario Diabetes Database (ODD). *Adjusted for age, sex, Charlson comorbidity and stroke type.

27,542Overall

7,749 34.00 (95%CI; 34.42-34.58)

19,793 35.51 (95%CI; 35.13-35.88)

16,308

DM 4,465 20.50 (95%CI;19.98-21.02)

11,843 20.87 (95%CI;20.55-21.19)

All Patients

No DM

1995 DM 51 1.9 2.477

711996

DMStatus

Median Waiting

Time (days)

DM 74

Men

Overall

Women

Overall

1.0

59

143

144

2.6

No DM 41 1.0 2.2150

2.2

OverallMen & Women

1997 DM 82 1.1 1.8

1.9

1.6

1.5

1.5

1.5

No DM 55 0.8 2.2

75

144

No DM 63 0.9 2.2

1998 DM 62 1.1 2.41.8

1.5No DM 70 1.0 2.2

70

141

1999 DM 74 1.3 2.01.7

1.6No DM

55 1.2 2.0

nFiscal Year Rate RateRate

1.07 (0.801.42)Odds Ratio Crude*(95% CI)

1.14 (0.711.82) 0.99 (0.691.43)

0.95 (0.711.27)Odds Ratio Adjusted*(95% CI)

1.00 (0.621.60) 0.99 (0.691.43)


12/14

Diabetes and Stroke

7.162

Diabetes in Ontario

infarction and comorbid illness. Rural residence, region of

residence, number of ambulatory care visits and having a regular

source of care were not significant predictors of subsequent

stroke admission.

Among hospitalized stroke patients, the characteristics of those

with and without DM were similar. Those with DM were

slightly younger and were more likely to be male, but the

differences were small (Exhibit 7.7). There were significant

differences in stroke type in those with and without DM, with

a higher proportion of ischemic stroke among those with DM

(94% vs. 89%, P


13/14

References1. Heart and Stroke Foundation of Canada. Statistics Canada, He

Canada, and Heart and Stroke Foundation of Canada, editors. H

disease and stroke in Canada, 1997.

2. Jamrozik K, Broadhurst RJ, Forbes S, Hankey GJ, Anderson CS. Predic

of death and vascular events in the elderly. The Perth CommuStroke Study. Stroke 2000; 31:863868.

3. Kuusisto J, Mykkanen L, Pyorala K, Laakso M. Non-insulin-depend

diabetes and its metabolic control are important predictors of strok

elderly subjects. Stroke 1994; 25:11571164.

4. Jorgensen H, Nakayama H, Raaschou HO, Olsen TS. Stroke in patiwith diabetes: the Copenhagen Stroke Study. Stroke 1994; 25:197719

5. Benson RT, Sacco RL. Stroke prevention: hypertension, diabetes, toba

and lipids. Neurol Clin 2000; 18:309319.

6. UK Prospective Diabetes Study Group. Intensive blood-glucose con

with sulphonylureas or insulin compared with conventional treatm

and risk of complications in patients with type 2 diabetes (UKPDS

Lancet1998; 352:837853.

7. Tuomilehto J, Rastenyte D, Jousilahti P, et al. Diabetes mellitus as a

factor for death from stroke. Prospective study of a middle a

Finnish population. Stroke 1996; 27:202205.

8. Hankey GJ, Jamrozik K, Broadhurst RJ, Forbes S, Burvill PW, AndeCS, Stewart-Wynne EG. Long-term risk of first recurrent stroke in

Perth Community Stroke Study. Stroke 1998; 29:24912500.

9. North American Symptomatic Carotid Endarterectomy T

Collaborators. Beneficial effect of carotid endarterectomysymptomatic patients with high-grade carotid stenosis. N Engl J M1991; 325:445453.

10. European Carotid Surgery Trialists Collaborative Group. MRC Europ

Carotid Surgery Trial: interim results for symptomatic patients w

severe (7099%) or with mild (029%) carotid stenosis. Lancet 19337:12351243.

11. Barnett HJM, Taylor DW, Eliasziw M, Fox AJ, Ferguson GG, Haynes

Rankin RN, Clagett GP, Hachinski VC, Sackett DL, et al. Benefi

carotid endarterectomy in patients with symptomatic moderatesevere stenosis. N Engl J Med1998; 339: 14151425.

12. Fabris F, Zanocchi M, Bo M, et al. Carotid plaque, aging and

factors. A study of 457 subjects. Stroke 1994; 25:11331140.

13. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity in

for use with ICD-9-CM administrative databases. J Clin Epi1992; 45613619.

14. Weiner JP, Starfield BH, Steinwachs DM, Mumford LM. Developm

and application of a population-oriented measure of ambulatory ccase-mix. Medical Care 1991; 29:452472.

15. Reid RJ, MacWilliam L, Verhulst L, Roos N, Atkinson M. Performance of

ACG Case-Mix system in two Canadian Provinces. Medical Care 239:3699.

16. Barrett-Connor E, Khaw K: Diabetes mellitus: An independent

factor for stroke. Am J Epidemiol1988; 128:116124.

very little difference in 30-day and one-year all-cause mortality

in those with and without DM (Exhibit 7.9).

Overall rates of carotid endarterectomy after stroke were low,

but similar in those with and without DM (Exhibits 7.10 and

7.11). The most significant finding was that rates of surgery

were about twice as high for men than for women regardless

of DM status.

ConclusionsDiabetes is an extremely powerful risk factor for stroke (it

increases the risk of stroke almost three-fold), in both men and

women and in every age group. Between 1995 and 1999, there

was a gradual decline in the risk of stroke hospitalization in

those with and without DM. This could be related to a decrease

in the incidence of stroke due to improved blood glucose

control, use of antithrombotic agents, or modification of other

stroke risk factors such as hypertension and hyperlipidemia.

Of note, there was an increase in the use of antihypertensive

and lipid lowering medications during the study time frame

(See Chapter 3: Drug Use in Older People with Diabetes).

However, it is difficult to draw firm conclusions given the

multifactorial etiology of stroke and the relatively short time

interval studied. Other potential explanations for the observed

decline in stroke hospitalization rates include changes in

admission thresholds for those with less severe strokes.

These analyses do not provide direct information on the

influence of diabetes on stroke severity. Persons with DM had

a slightly longer length of stay, and were more likely to be

either discharged to complex continuing care facility or die

within 30 days of stroke, regardless of stroke type. While this

could indicate greater stroke severity in those with DM, other

explanations include a greater frequency of post-stroke

complications or other comorbid illness. The finding that

adjusted 30-day and 1-year all cause mortality after stroke

were not increased in those with DM argues against major

differences in stroke severity based on DM status.

Post-stroke carotid endarterectomy rates were similar in those

with and without DM. It is surprising that women were only

half as likely as men to undergo carotid endarterectomy, even

in the presence of DM where stroke risks are similar in women

and men. Data sources with more detailed clinical information

are needed to determine the prevalence of moderate to severe

carotid stenosis in men and women with and without DM and

the appropriate rates of carotid endarterectomy in these

populations.

Overall, these analyses confirm that stroke is a common and

serious complication of DM. Further study is needed to determine

whether local initiatives to improve DM care will result in

significant reductions in stroke risks or improved stroke outcomes.

Practice Atlas

7.163


14/14


DM Chapter7

Documents

Transcript of DM Chapter7